Feasibility of Ultraviolet Light Emitting Diodes as an Alternative Light Source
3 Lanfang H. Levine and Jeffrey T. Richards
4 Dynamac Corporation, Kennedy Space Center, FL
5 Robert Soler and Fred Maxik
6 Lighting Science Group Corporation, Satellite Beach, FL
7 Janelle Coutts
8 Department of Chemistry, University of Central Florida, Orlando, FL
9 Raymond M. Wheeler
10 Engineering Directorate (NE-S), Kennedy Space Center, FL
1 1 ABSTRACT
12 The objective of this study was to determine whether ultraviolet light emitting diodes
13 (UV-LEDs) could serve as an alternative photon source efficiently for heterogeneous
14 photocatalytic oxidation (PCO). An LED module consisting of 12 high-power UV-A
15 LEDs was designed to be interchangeable with a UV-A fluorescent black light blue
16 (BLB) lamp in a Silica-Titania Composite (STC) packed bed annular reactor. Lighting
17 and thermal properties were characterized to assess the uniformity and total irradiant
18 output. A forward current of (If) 100 mA delivered an average irradiance of 4.0 mW
19 cm' 2 , which is equivalent to the maximum output of the BLB, but the irradiance of the
20 LED module was less uniform than that of the BLB. The LED- and BLB-reactors were
21 tested for the oxidization of 50 ppm v ethanol in a continuous flow-through mode with
22 0.94 sec space time. At the same irradiance, the UV-A LED reactor resulted in a lower
23 PCO rate constant than the UV-A BLB reactor (19.8 vs. 28.6 nM CO 2 sec' 1 ), and
24 consequently lower ethanol removal (80% vs. 91%) and mineralization efficiency (28%
25 vs. 44%). Ethanol mineralization increased in direct proportion to the irradiance at the
26 catalyst surface. This result suggests that reduced ethanol mineralization in the LED-
27 reactor could be traced to uneven irradiance over the photocatalyst, leaving a portion of
28 the catalyst was under-irradiated. The potential of UV-A LEDs may be fully realized by
29 optimizing the light distribution over the catalyst and utilizing their instantaneous “on”
30 and “off’ feature for periodic irradiation. Nevertheless, the current UV-A LED module
3 1 had the same wall plug efficiency (WPE) of 1 3% as that of the UV-A BLB. These results
32 demonstrated that UV-A LEDs are a viable photon source both in terms of WPE and
33 PCO efficiency.
36 Hg-vapor lamps are common UV sources for photocatalysis but create safety and environmental
37 concerns because they contain Hg; furthermore they have a relatively short life span. This paper
38 demonstrated that the UV-A LED is a viable alternative to the Hg-vapor lamps without
39 sacrificing PCO efficiency if the design of the LED arrays is improved to increase the irradiant
40 uniformity. The use of LEDs could eliminate hazardous Hg wastes and extend photocatalysis
41 application to places requiring more compact and robust air purification solutions.
43 The ability of titanium dioxide (Ti02)-assisted photocatalytic oxidation (PCO) to decompose
44 (mineralize) a broad range of organic contaminants into CO 2 and H 2 O at room temperature has
45 attracted attention for various environmental applications. This technique has been investigated
46 as an alternative or complimentary method for air contaminant control 1 ' 7 as well as a means for
47 treating water and wastewater. 7 The Ti02-catalyzed PCO process typically requires a light source
48 with a wavelength less than 388 nm. Mercury vapor lamps such as UV-A black lights and UV-C
49 germicidal lamps have been widely used in laboratory and commercial PCO systems (e.g.
50 GENESIS AIR PHOTOCATALYST GAP™, Ultra Sun Technologies, and Mazyck
51 Technologies). But these lamps contain trace amounts of toxic mercury, there are fragile, and
52 have a relatively short life span (<12,000 hrs). Mercury is a highly toxic and controlled
53 substance, and it is increasingly becoming controlled or banned by government safety and
54 environmental regulators. Although there are non-mercury lamps, e.g., microwave generated
55 UV sources, the lamps are driven by magnetrons that must withstand long duty cycles and the
56 microwaves must be contained for safety purpose.
58 On the other hand, light emitting diodes (LEDs, semiconductor-based lighting devices) are
59 compact, reliable, and long lasting devices. LEDs are driven by direct current, can accommodate
60 faster switching, and do not contain toxic Hg. They are entering the market of various
61 illumination applications with an unprecedented speed. Since the development of first
62 commercial visible LEDs in 1968, the light output from a single device has increased by a factor
63 of 20 per decade, while the price in US dollar per lumen has declined by a factor of 10 per
64 decade. 8 White-light LEDs are now surpassing the efficiency of linear fluorescent and compact
65 fluorescent lamps. 8 Ultraviolet light emitting diodes (UV-LEDs) have been commercially
66 available since 2003. Currently, UV-A LEDs have a life expectancy of 50,000 hrs at L50, about
67 5 times that of Hg-vapor lamps. Naturally they have been considered as an alternative light
68 source for photocatalysis for both gaseous 9 ’ 10 and aqueous applications. 1 1,12 However, most of the
69 PCO-studies to date were conducted with low-power LEDs of varied wavelengths including
70 those outside of the TiC >2 action spectrum (e.g. 395 and 405 nm 11,12 ) in different reactors. Chen
71 et al. studied photocatalytic degradation of percholoroethylene (PCE) in a rectangular steel gas-
72 phase reactor irradiated by 375 nm UV LEDs (16 Nichia LEDs with 1.0 mW power output) and
73 found only 43% degradation of PCE in the LED reactor, while there was 90% degradation in a
74 UV-A black light reactor. 9 This result seemed to imply that LEDs are less effective than the
75 black light. Ciambelli et al. investigated the photocatalytic breakdown of benzene in a lab scale
76 fluidized bed reactor irradiated by two or four UV (365-nm) LED modules (Nichia Corporation)
77 and showed a 27% conversion of benzene into CO 2 at 80 °C. 10 Although these studies proved
78 that UV LEDs have promise as a light source, no data were provided regarding their PCO
79 performance relative to mercury-vapor lamps at the same irradiance, their actual power use
80 efficiency, or issues related to the LED integration into PCO reactors. The objective of this study
81 was to design an LED PCO reactor and compare the performance of state-of-the art UV-A LEDs
82 to that of a Hg-vapor UV-A fluorescent black light for low temperature PCO degradation of
83 organic contaminants. '
86 EXPERIMENTAL METHODS
87 Light Sources
88 An 8-W UV-A (F8T5) fluorescent black light blue (BLB) lamp from Philips was used as the
89 control light source. The BLB lamp dimensions were 15.6 mm D x 304.8 mm L and the irradiant
90 output was 2.5 mW cm' 2 at a 25.4 mm distance. The LEDs used for the study were high-power
91 chip-type UV-A LEDs (model NCSU033B) from Nichia Corporation, having a peak wavelength
92 of 365 nm, a spectrum half-width of 9 nm, and irradiance angle of 120 degrees. The optical
93 output of a single LED was 325 mW at a forward current of 500 mA and voltage of 3.8 V (i.e.,
94 1.9 Watts).
96 To conduct direct comparison of PCO efficiency with UV-A BLB, the LED source was designed
97 based on three criteria: 1) interchangeability with the BLB in the same PCO reactor; 2) similarity
98 of irradiance profile between the LED module and the BLB (i.e. isotropic); and 3) a wide range
99 of irradiances, including one equivalent to that of the 8-W BLB. Based on this, we designed an
100 LED assembly to simulate the geometry of the linear fluorescent bulb. The placement of
101 individual LEDs was determined by modeling (Figure 1). Uniformity of the irradiance increased
102 as the distance between LEDs decreased (Figure la). Thus an assembly with densely populated
103 LEDs (e.g., 15 mm vs. 30 mm spacing between LEDs) would provide more uniform irradiance,
104 but the power consumption and initial investment (number of the LEDs) would increase
105 accordingly. In addition, greater numbers of LEDs create thermal management challenge, where
106 lower operating temperatures are preferred to maintain the long operating life of the LEDs. The
107 coefficient of temperature increase per unit electric power input is dependent upon the thermal
108 resistance of the LED system (e.g., Rj a = 35 °C/W with Nichia’s standard circuit board), density
109 of the LED placement, and factors such as ambient temperature. The radial irradiance profile of
110 the LED assembly was modeled based on four linear LED arrays evenly placed every 90-degrees
111 in a 360-degree arrangement. The model examined the effect of viewing angle and distance
112 between the light source and the object to be irradiated (Figure lb). Since the object to be
113 irradiated was the photocatalyst packed in an annular reactor (see details in PCO reactor design),
114 optimal diameter (0) of the quartz sleeve separating the light source and the catalyst was
1 15 subsequently determined by this modeling exercise. Although increasing the diameter enhanced
116 the uniformity of light distribution, the radiant flux per unit area (E) decreased approximately
117 following an inverse-square law (Figure lb). Figure 1 suggests that 20-mm spaces between
118 LEDs, and a 25-mm diameter would give satisfactory uniformity and sufficient intensity.
119 Consequently, twelve LEDs were mounted on a 15.6-mm diameter aluminum rod in four linear
120 arrays, three LEDs in each array with a space of 20 mm between the LEDs as shown in Figure 2.
122 Figure 1 and Figure 2 here
123 c ' Photocatalytic Oxidation (PCO) Reactor
124 Silica-Titania Composite (STC) pellets (2x6 mm) from Sol-gel Solutions, LLC (Gainesville,
125 FL) served as the photocatalyst. The STC has the same porosity (30 - 40 A) and Ti02 loading (4
126 g Degussa P-25 in 100 mL silica precursor, tetraethyl orthosilicate) as that used by Stokke and
127 Mazyck. 13 An annular reactor shown in Figure 2 was used to carry out this study because of its
128 simplicity and efficiency in utilizing traditional linear fluorescent lamps. It consisted of two
129 concentric cylinders, with an annulus formed between an aluminum housing and a quartz sleeve.
130 The light source was inserted in the middle of the quartz sleeve, while STC pellets were packed
131 in the annulus. Two design parameters were optimized: 1) the diameter of quartz sleeve that
132 determines the distance between the photocatalyst and light sources; and 2) the annulus size.
133 The former was especially critical with LEDs as the light source. The effect of the quartz
134 sleeve’s diameter was illustrated in Figure lb. The annulus space that determines the thickness
135 of the catalyst bed should be small enough to ensure that photons emitted from the light source
136 reach all catalyst surfaces uniformly and large enough to allow reproducible packing of STC
137 pellets. Results of a preliminary light transmittance measurement showed that STC pellets in an
138 annulus of 8 mm used by Stokke and Mazyck 13 attenuated 97% of a UV-A BLB light. This
139 suggested the annulus size of the reactor should be further reduced. Key parameters of the bench-
140 scale test reactor used in this study and that used by Stokke and Mazyck 13 are listed in Table 1
141 for comparison.
143 Table 1 here
144 Light Source Characterization
145 Spectral quality and quantity of the light sources were assessed to determine an optimal UV-A
146 BLB for the study and the driving current required for the LED assembly to achieve an
147 equivalent irradiance. Measurements were conducted outside the PCO reactor in a dark room
148 using a spectroradiometer (OL754C, Optronics Laboratories, Orlando, FL). The light source
149 (either BLB or LED assembly) was centered inside a quartz sleeve of the same dimension as that
150 used in the reactor (Table 1) without the catalyst around it and placed directly above the
151 integrating sphere of the spectroradiometer. A light attenuation aperture of 12.7-mm in diameter
152 was placed before the integrating sphere that has an opening of 31.8 mm in diameter. This setup
153 (Figure 3) measured the irradiance immediately at the surface of the catalyst bed. For the LED
154 module, measurements were taken every 5 mm along the lateral direction both directly opposite
155 to one of four LED arrays (designated as angle 0 degree) as shown in Figure 3 and between two
156 linear LED arrays (designated as angle 45 degree). Irradiant output of the LED assembly and
157 LED die’s temperature were measured at a range of forward current between 30-500 mA. The
158 BLB was measured at three positions along the axis of the lamp.
160 Figure 3 here
161 PCO Tests
162 Performance of the UV-A BLB- and the LED-irradiated annular reactors was evaluated for
163 oxidation of ethanol in an experimental setup (Figure 4) that allowed precise control of
164 experiment variables and continuous monitoring of the PCO reaction. The setup consisted of: 1)
165 a Kin-Tek air generator (model 49 1M, La Marque, TX) for supplying a simulated contaminant
166 air containing 50 ppm v ethanol (EtOH) and 72% relative humidity at 25 °C; 2) a PCO reactor
167 packed with 14.6 g of STC pellets to a bed height of 60 mm; 3) two mass flow controllers for
168 controlling the flows to the PCO reactor and CO 2 analyzer; 4) temperature sensors for the
169 reactor’s inlet and outlet as well as for room temperature; 5) humidity sensors for the reactor’s
170 influent and effluent air; 6) a CO 2 analyzer for the reactor effluent; 7) a sample stream selecting
171 valve; and 8) a gas chromatograph (ThermoFinnigan, Austin, TX) equipped with a flame
172 ionization detector (GC/FID) and a HP Plot Q capillary column (30 m x 0.32 mm, 20 pm depth
173 of film).
175 Figure 4 here
177 All tests were carried out in a flow-through mode with an uninterrupted 2 L min' 1 air flow
178 containing 50 ppm v EtOH under continuous illumination. Each test was repeated three times.
179 Both influent and effluent were sampled alternately every 8.45 min and analyzed for ethanol and
180 any oxidation intermediates by GC/FID. The effluent was also directed to a CO 2 analyzer to
181 determine the production of CO 2 , the complete mineralization product. CO 2 concentration was
182 recorded every minute. The reactor was maintained at 25 °C via forced air convection using a
183 heat sink attached to the PCO reactor. The STC pellets were pre-conditioned with 72% RH,
184 VOC-free air under continuous illumination. Each test began with the addition of ethanol to the
185 air stream and continued for 21 hours, followed by regeneration with humidified, VOC-free air
186 and continuous illumination. The same batch of STC catalyst was used for all runs. Completion
187 of the regeneration was indicated by no detectable organic species and only baseline-level CO 2 in
188 the effluent.
190 PCO Efficiency, Kinetics and Photonic Efficiency
191 PCO performance was quantified in terms of EtOH removal and mineralization efficiency (Xa).
192 The former is a measure of the total removal of the test VOC, whether it is removed by
193 adsorption or oxidation, while the latter is a measure of the complete oxidation of EtOH to CO 2 .
194 These values were calculated using equations 1 and 2, respectively. Co and Ceioh are the influent
195 and effluent EtOH concentrations, and AC car bon dioxide is the CO 2 generated from the PCO process.
196 The rate of photocatalytic oxidation of ethanol was determined based on the formation of CO 2
197 instead of the disappearance of ethanol to prevent overestimation due to the adsorption of EtOH
198 to the photocatalyst. Cumulative CO 2 concentration was plotted against time, a linear
199 relationship between the concentration and time suggested zero-order kinetics. The slope gave
200 rise to the PCO rate (r). PCO photonic efficiency (^) was calculated as the ratio of the
201 photocatalytic degradation rate to the incident photon flux (eq 3).
202 EtOH removal = (Co-CeiohVCo (1)
203 Xa AC C arbon dioxide /2 X Co (2)
204 § = Rate of reaction (M sec‘')/rate of photon incident (mol sec" 1 ) (3)
206 RESULTS AND DISCUSSION
207 Spectral Quality and Quantity of the Light Sources
208 Photon flux, or irradiance, on the catalyst surface is one of the most important factors affecting
209 photocatalytic oxidation efficiency. The LED assembly was extensively characterized in order to
210 assess its irradiance uniformity and the required driving current for the LED module to provide
211 an optical output similar to that of an 8-W BLB. Initial scans of four UV-A fluorescent black
212 lights from GE, Eiko, Philips, and Sylvania demonstrated that the GE and Eiko lamps are similar
213 in their spectra and intensity. The Philips lamp ranked the highest in irradiance among the four
214 lamps examined, while the Sylvania lamp had the lowest irradiance and a very broad peak.
215 Hence, the Philips brand lamp was used in this study. Relative to the UV-A LED, the UV-A
216 BLB had a broader peak (354-388 nm) centered at 365 nm and an additional peak at 405 nm that
217 is out of the TiC >2 action spectrum (<388 nm) (Figure 5). The LED spectrum peak was narrower
218 (357-378 nm) and all of the radiation fell within the Tick’s action spectrum. Furthermore, the
219 spectra of adjacent LED linear arrays (LED Array 1 and Array 2) were identical (Figure 5).
221 Figure 5 here
223 Lateral irradiance profile of the LED assembly was measured every 5 mm from the first LED
224 along the lateral axis directly opposite to one of the 4 arrays (angle 0 degree) as well as opposite
225 to the space between two arrays (angle 45 degree) at the forward current of 100 mA. Irradiance
226 from the LED assembly was not less uniform than desired (Figure 6). The lowest intensity was
227 about 55% of the peak intensity, occurring directly between two LEDs in an individual array.
228 The average irradiance (E) at angle 0 degree was 6.02 mW cm' 2 with 2.49 mW cm’ 2 at angle 45
229 degree, resulting in a mean of 4.25 mW cm' 2 for the assembly. The overall mean irradiance for
230 the LED module was 70% of the predicted value (Figure 6b). The discrepancy could be
231 explained by the directionality (120 degree) of the LED radiation and how the light was
232 measured (Figure 3). The combination of the small sensor aperture (12.7 mm) and the close
233 distance (approximately 8 mm) between the sensor and the light source prevented some of the
234 photons from adjacent LEDs from entering the integrating sphere (Figure 3a). As a result, the
235 measured value was underestimated comparing with that obtained in the absence of the
236 attenuation aperture (Figure 3b). Nevertheless, the opacity of STC pellets packed immediately
237 outside of the quartz sleeve in a working PCO reactor would act as the attenuation aperture and
238 prevent the photons outside the radius of 12.7 mm aperture from reaching the catalyst located
239 where the light sensor was placed. We believed that the measured value was a more accurate
240 representation of the light level the catalyst would intercept rather than the predicted value. In
241 contrast, the 8-W UV-A BLB lamp from Philips measured in the same way showed a uniform
242 irradiance of 4.0 ± 0.2 mW cm' 2 along the both radial and lateral axes.
243 Figure 6 here
245 Furthermore, the light output of the LED assembly was also measured at If between 30 to 500
246 mA. As with the individual LEDs, the irradiance of the LED assembly was directly proportional
247 to the driving current in the range of 30 to 500 mA (E = 0.0449If-0.2235, R 2 = 0.9999).
248 Consequently, a nominal 100 mA driving current for the LED assembly delivered the maximal
249 irradiance of an 8-W fluorescent lamp.
250 Thermal Characteristics of the LED Assembly
251 LED life span can vary according to environmental and design related factors. Although it is
252 largely determined during the engineering phase of an LED lighting design, overdriving an LED
253 assembly will decrease its life span if thermal management is inadequate. In order to assess the
254 effectiveness of our heat management strategy and to determine the upper limit of driving current
255 (and hence the light output) for the assembly, the temperature of each LED in the assembly was
256 measured at three driving currents (Table 2). The LED temperatures (Tj) were calculated based
257 on the thermal resistance from the LED die to the measuring point being 7 °C W' 1 . Results
258 showed that a linear relationship between the driving current and measured solder temperature
259 (T s ) or calculated junction temperature (Tj), that is, Tj = 0.0857If + 25.4 (R 2 = 0.9999). From
260 this, the maximal allowable driving current was determined to be 870 mA to operate the LEDs
261 below the manufacturer’s recommended maximum Tj of 100 °C. Because the LED used in this
262 study was rated for a maximum forward current 700 mA, the assembly consisting of 12 LEDs
263 electrically strung in two parallel series should allow for a maximum of 1400 mA and result in a
264 light output of 62.6 mW cm' 2 based on the established relationship between the irradiance and
265 forward current (E = 0.0449If-0.2235). It was determined that the LED assembly had a greater
266 light output potential (62.6 mW cm' 2 ) than that the current thermal management strategy could
267 deliver (38.9 mW cm' 2 ). That is, from the thermal perspective, the assembly can only fulfill 62%
268 of its light output potential. This is primarily due to the design constraints for this first
269 generation LED module to be directly comparable with linear fluorescent lamps. Four linear
270 LED arrays were mounted on a small aluminum rod; thermal energy (e.g., 12 W at 500 mA
271 driving current) had to be conducted to the ends for convective dissipation.
273 Table 2 here
274 PCO Efficiency of the BLB and LED-irradiated Reactors
275 The PCO reactor effluent was found to consist of ethanol (VOC contaminant), acetaldehyde
276 (oxidation intermediate), and carbon dioxide (final oxidation product). Acetaldehyde (ACD)
277 was the only quantifiable intermediate in the effluent as indicated by the lack of any other peaks
278 in the GC chromatograms (data not shown). The UV-A BLB-irradiated reactor generated
279 effluent ethanol and acetaldehyde profiles (Figure 7b) similar to those reported for methanol
280 oxidation. 14 Upon the initiation of ethanol-contaminated air flow, effluent ethanol concentration
281 remained very low (2% of the feed) for the first three hours, increased at an accelerated rate
282 between 3 and 10 hrs, and continuously crept upwards even after 10 hrs. This initial lag time for
283 ethanol was attributed to the adsorption of ethanol by STC pellets. In contrast, there was a very
284 short (less than 30 min) or no initial lag time for ACD and C0 2 , respectively. The concentration
285 of ACD and C0 2 in the effluent increased steeply upon the addition of feed contaminant,
286 suggesting low and/or no affinity of ACD and C0 2 to the STC. The concentration of C0 2
287 approached a plateau or a steady state between 5 and 10 hrs, but that of ACD and ethanol
288 reached somewhat steady state only after 10 hr. Therefore, the time period between 10 and 20
289 hrs was considered as the “pseudo-steady state.” The time course profiles of effluent ethanol,
290 ACD, and C0 2 from the UV-A LED reactor (Figure 7a) resembled those of the BLB reactor in
291 general shape, but differed in slope and concentration level at the pseudo-steady state.
293 Figure 7 and Table 3 here
295 Mineralization of ethanol in both reactors followed zero-order kinetics and had a rate constant of
296 19.8 and 28.6 nM C0 2 sec' 1 for the LED and BLB, respectively. The average concentration of
297 effluent components at this time period was used to assess the PCO efficiency in terms of ethanol
298 removal and mineralization (Table 3). Compared with the UV-A BLB reactor, the LED reactor
299 had a lower effluent ACD and C0 2 but higher EtOH, which translated into lower EtOH removal,
300 mineralization, POC rate and photonic efficiency than the UV-A BLB reactor (Figure 8b through
301 e). The results do not necessarily indicate that the LEDs were a less effective light source,
302 bearing in mind that the LED module’s irradiance was not as uniform as the BLB; some of the
303 catalyst was irradiated by less than the average irradiance, which may have accounted for the
304 reduced C0 2 and ACD in the effluent.
305 The effect of irradiance on PCO efficiency was subsequently examined in the LED reactor.
306 Increasing the irradiance at the catalyst surface reduced effluent ACD and EtOH and increased
307 CO 2 production (Figure 8a). There was a linear relationship between both the mineralization
308 efficiency and rate constant and the irradiance (Figure 8c & d). The influence of increasing
309 irradiance on the percentage of ethanol removal (Figure 8b) was not as pronounced as for the
310 mineralization. This is not surprising and is attributable to the STC’s unique property of high
311 physical adsorptivity for polar compounds and photocatalytic activity. However, increasing the
312 irradiance from 4.0 to 17.7 mW cm" 2 decreased the photonic efficiency (§) by 33%. The linear
313 relationship between mineralization and irradiance suggested that an irradiance of 7.6 mW cm' 2
314 from the LED module would reach the same mineralization as that of the BLB at its full
315 intensity. In other words, the LED reactor used in this study could achieve the same PCO
316 efficiency for ethanol as an 8-W UV-A FL when it is operated at a forward current (I F ) of 170
317 mA (i.e., a power input of 3.4 W). These values could be reduced by a more uniform irradiance
3 1 8 over the entire surface of the photocatalyst.
320 Figure 8 here
322 Power Use Efficiency
323 Power use efficiency of a light source for PCO encompasses both the electric-irradiant
324 efficiency, or wall plug efficiency (WPE), and the PCO efficiency. WPE (defined as the
325 percentage of irradiant output per electrical input) of the light sources used in this study is shown
326 in Table 4. It is apparent that the WPE of the UV-A LED and UV-A FL were comparable, 17%
327 as the manufacturer specified and 13% as measured in this study. It is interesting to note that the
328 WPE efficiency is higher for longer wavelength LEDs. For example, the same type of LED with
329 spectrum centered at 385 nm has a WPE of 21.6%, representing a 25% increase from that of the
330 365 nm LEDs. However, it is not known whether the gain in WPE would be offset by the
331 potential loss in PCO efficiency since 385-nm LEDs approach the upper limit of the Ti02 action
332 spectrum. In terms of PCO efficiency, LEDs in the current design were slightly less efficient
333 when compared to the UV-A FL, but the gap could be closed if a more uniform irradiance over
334 the catalyst is achieved. In addition, PCO efficiency of LEDs could be enhanced by exploitation
335 of its instantaneous “on” and “off feature for periodic irradiation. It was previously
336 demonstrated that photonic efficiencies for decomposition of o-cresol by a UV/TiCL process in a
337 slurry reactor under controlled periodic illumination of LEDs was higher than that under
338 continuous illumination. 12 As a result, the electric energy required for degradation of the same
339 amount of contaminants decreased significantly by using periodic irradiation.
341 Table 4 here
344 This is the first report of a direct comparison between UV-A LED and UV-A BLB as PCO light
345 sources under similar irradiance. Challenges encountered in achieving uniform LED irradiance
346 over the photocatalyst while maintaining the power use efficiency. Increasing the density of
347 LEDs could no doubt enhance the uniformity of the irradiance, but it would also increase the
348 initial cost of a PCO reactor and the burden of heat management if high-power LEDs are used.
349 The results from our LED reactor suggest that the number of LEDs per unit area may actually be
350 reduced because of the facts that the LED assembly could deliver up to 38.9 mW cm' 2 and an
351 irradiance of 17.7 mW cm' 2 resulted in a 97% EtOH removal and an 86% mineralization. It
352 became clear that different design strategies should be considered depending upon the type of
353 UV-A LEDs 15 to be used, for instance, a higher density of low-power LEDs (<10 mW) or high-
354 power LEDs (> 1 00 mW) coupled with light dispersion. Typical approaches for light dispersion
355 in lab-scale reactors include a) coupling LEDs to light transmitting optical fibers coated with a
356 thin film of photocatalysts, and b) using a waveguide through which light travels and emits from
357 sides into surrounding catalysts. Although these approaches have the advantage of transferring
358 small area of a LED’s illumination to a much greater surface area, each has its own drawbacks.
359 The former has limited applications to thin films of catalysts because of TiCVs opacity, while the
360 later creates a gradient of irradiance along the waveguide. A balance between side-emitting and
361 transmitting must be struck to achieve uniform side emission intensity over reasonable lengths.
362 We are currently working on screening light conduit materials & LED-light conduit coupling
363 techniques for effective dispersion of high-power LED’s radiation.
365 PCO efficiency in terms of ethanol removal and mineralization was greater in the UV-A BLB
366 reactor than in the UV-A LED reactor at the same average irradiance (4.0 ± 0.2 mW cm' 2 ).
367 Irradiance level and uniformity over the catalyst was found to have a great impact on the PCO
368 efficiency. PCO efficiency increased linearly as the irradiance over the surface of catalyst
369 increased in the range tested (4-18 mW cm' 2 ). We estimated that the LED reactor used in this
370 study could achieve the same ethanol mineralization as a 8-W UV-A BLB when it was operated
371 at forward currents (If) of 170 mA, which corresponded to a power input of 3.4 W and an
372 irradiant output of 7.6 mW cm' 2 . These values are expected to be lower as uniform irradiance
373 and/or periodic irradiation are implemented. The results proved that LEDs are a viable photon
374 source both in terms of PCO efficiency and wall plug efficiency. Continuing efforts in the
375 following areas will strengthen this conclusion: 1) improvements in the design of the LED-PCO
376 reactor for a higher fidelity estimate of power use efficiency; 2) investigation of the trade-off
377 between PCO efficiency and electric-irradiant efficiency by using longer wavelength LEDs (e.g.
378 385 nm instead of 365 nm); 3) using visible light responsive catalysts to take advantage of the
379 higher quantum efficiency of longer wavelength LEDs.
381 The work was conducted under the auspices of Life Science Support Contract and the first part
382 of a Kennedy Space Center Innovative Partnership Program (IPP) funded project to highlight the
383 partnership with Lighting Science Groups Corporation (LSGC). Authors are extremely grateful
384 to Dr. David Mazyck of the University of Florida for donating the photocatalyst. The authors
385 would like to thank Mr. Lawrence L. Koss for his invaluable assistance with the PCO test bed
386 construction by making customized parts and Opto 22 data logging. We would also like to
387 extend our appreciations to Mr. J. Schellack and Mr. D. Johnson of LSGC for constructing the
388 LED assembly and KSC prototype shop personnel for fabricating the bench scale PCO reactor.
1. Hoffmann, M.R.; Martin, S.T.; Choi, W.; Bahnemannt, D.W. Environmental applications of
semiconductor photocatalysis; Chem. Rev. 1995, 95, 69-96.
2. Mo, J.; Zhang, Y.; Xu, Q.; Lamson, J.J.; Zhao , R. Photocatalytic purification of volatile
organic compounds in indoor air: a literature review; Atmos. Environ. 2009, 43, 2229-2246.
3. Hodgson, A. T.; Sullivan, D. P.; Fisk, W. J. Evaluation of ultra-violet photocatalytic
oxidation (UVPCO) for indoor air applications: conversion of volatile organic compounds at
low part-per-billion concentrations; Lawrence Berkeley National Laboratory, 2005, LBNL-
4. Zhao, J.; Yang, X. D. Photocatalytical oxidation for indoor air purification: a literature
review; Build, and Environ. 2003, 38, 645-654.
5. Tompkins, D. T.; Lawnicki, B. J.; Zeltner, W. A.; Anderson, M. A. Evaluation of
photocatalysis for gas-phase air cleaning - Part 1 : process, technical and sizing
considerations; ASHRAE Trans. 2005, 1 1 1(2), 60-84.
6. Tompkins, D. T.; Lawnicki, B. J.; Zeltner, W. A.; Anderson, M. A. Evaluation of
photocatalysis for gas-phase air cleaning - Part 2: economics and utilization; ASHRAE Trans.
2005, 111 (2), 85-95.
7. Kwon, S.; Fanb, M.; Cooper, A.T.; Yang, H. Photocatalytic applications of micro- and nano-
TiC >2 in environmental engineering; Crit. Rev. in Environ. Sci. and Tech. 2008, 38(3), 197-
8. Steele, R.V. The story of a new light source; Nature Photonics 2007. 1, 25-26.
9. Chen, D. H.; Ye, X.; Li, K. Oxidation of PCE with a UV LED photocatalytic reactor, Chem.
Eng. Technol. 2005, 28, 95-97 .
10. Ciambelli, P.; Sannino, D.; Palma, V.; Vaiano, V.; Mazzei, R. S.; A step forwards in ethanol
selective photo-oxidation; Photochem. Photobiol. Sci. 2009, 8(5), 699-704.
11. Wang, W. Y.; Ku, Y. Photocatalytic degradation of Reactive Red 22 in aqueous solution by
UV-LED radiation; Water Res. 2006, 40, 2249 - 2258.
12. Chen, H. W.; Ku, Y.; Irawan, A. Photodecomposition of o-cresol by UV-LED/Ti02 process
with controlled periodic illumination; Chemosphere 2007, 69, 184-190.
13. Stokke, J. M.; Mazyck, D. W. Effect of catalyst support on the photocatalytic destruction of
419 VOCs in a packed-bed reactor; Proceeding of 37th International Conference on
420 Environmental Systems, 2007 , SAE paper 2007-01-3138.
421 14. Stokke, J. M.; Mazyck, D. W. Photocatalytic degradation of methanol using silica-titania
422 pellets: Effect of pore size on mass transfer and reaction kinetics; Environ. Sci. & Tech.
423 2008 , 42(10), 3803-3813.
424 15. Sandhu, A. The future of ultraviolet LEDs; Nature Photonics 2007, 1, 38.
426 About the Authors
427 Lanfang H. Levine and Jeffrey T. Richards are senior research chemists with Dynamac
428 Corporation at Kennedy Space Center, FL 32899. Robert Soler and Fred Maxik are both
429 electrical engineers and serve as director of electrical engineering and chief scientific officer,
430 respectively for Research and Development of Lighting Science Group Corporation at 1227 S.
431 Patrick Dr. Satellite Beach, FL 32937. Janelle Coutts is a graduate student from the University of
432 Central Florida and currently a chemistry intern with Dynamac Corporation at Kennedy Space
433 Center. Raymond M. Wheeler is a plant scientist with NASA at Kennedy Space Center, FL
434 32899. Please address correspondence to Lanfang FI. Levine, Space Life Sciences Laboratories,
435 Mail Code: DYN-3, Kennedy Space Center, FL 32899, USA; phone: 321-861-2931; e-mail:
436 lanfang. h. levine@nasa. gov .
440 , Table 1 . Comparison of test reactors
Reactor used in Stokke and
Annular reactor in this
Reactor Housing ID (mm)
Quartz Sleeve OD (mm)
Annulus Space (mm)
Catalyst Bed Height (mm)
Bed Volume (cm 3 )
30, 60, 90
Controlled to 25 °C
Table 2. LED solder temperature (T s ) and dice temperature (Tj) as a function of If
If (mA) for the LED
Measured T s
Table 3. Effluent composition and PCO efficiency at pseudo-steady state from the BLB and
LED reactors at the same irradiance. Values represent the mean
Influent ethanol concentration was 5 1 ± 0.3 ppm v .
(± s.e.) between 10 and 20 hrs.
Average E, mW cm'
Effluent EtOH, ppm v
Effluent ACD, ppm v
Effluent CO 2 , ppm v
EtOH Removal (%)
91.0 ± 1.3
Table 4. Wall plug efficiency (WPE) of the light sources
Electric Input (W)
Optical Output (W) Specified
WPE (%) Specified
Optical Output (W) Measured
WPE (%) Measured
451 FIGURE CAPTIONS
453 Figure 1. Effect of LED spacing and distance away from the LED (i.e. Vi 0) on: (a) lateral
454 irradiance uniformity and (b) radial irradiation uniformity.
456 Figure 2. A 3-D model of the annular reactor shown with the LED light source.
458 Figure 3. Schematic of the setup for light source characterization, illustrating the effect of
459 aperture size on the amount of photons from the adjacent LEDs entering the integrating sphere.
461 Figure 4. Schematic of a bench-scale PCO test bed where the objects are not to scale.
463 Figure 5. Spectra of UV-A LEDs and fluorescent black lights.
465 Figure 6. Irradiance profiles of the LED assembly determined at Ip=100 mA: (a) lateral and
466 radial profiles on the outer surface of the quartz sleeve (OD 28 mm) where the photocatalyst is
467 located; (b) comparison between measured average and model-predicted irradiance.
469 Figure 7. Time-course of the effluent composition during STC-catalyzed oxidation of ethanol
470 in the (a) UV-A LED and (b) UV-A BLB reactors at the same irradiance of 4 mW cm' 2 . CO 2
471 concentration was recorded every minute and appears to be affected by the sample stream valve
472 position giving two parallel trend lines.
474 Figure 8. Effect of the average irradiance over the catalyst surface on STC-catalyzed PCO in
475 the LED reactor: (a) reactor effluent composition at the pseudo-steady state, PCO efficiency in
476 terms of (b) ethanol removal, (c) ethanol mineralization, and (d) PCO rate constant, and (e)
477 photonic efficiency.
Figure 1. Effect of LED spacing and distance away from the LED (i.e. 'A 0) on: (a) lateral
irradiance uniformity and (b) radial irradiation uniformity.
491 Figure 2. A 3-D model of the annular reactor shown with the LED light source.
w/ 31.8 mm aperture
Figure 3. Schematic of the setup for light source characterization, illustrating the effect of
aperture size on the amount of photons from the adjacent LEDs entering the integrating sphere.
Carbon Dioxide Analyzer
0.65 psi 56.2%
PT 2 RH2
< Viaii OM Mod
PCO System On
Figure 4. Schematic of a bench-scale PCO test bed where the objects are not to scale.
503 Figure 5. Spectra of UV-A LEDs and fluorescent black lights.
504 Distance from one LED to the other (mm)
506 Figure 6. Irradiance profiles of the LED assembly determined at I F =100 mA: (a) lateral and
507 radial profiles on the outer surface of the quartz sleeve (OD 28 mm) where the photocatalyst is
508 located; (b) comparison between measured average and model-predicted irradiance.
(a) UV-A LED
(b) UV-A BLB
Time of Continuous VOC Feed and Illumination (hr)
Figure 7. Time-course of the effluent composition during STC-catalyzed oxidation of ethanol
in the (a) UV-A LED and (b) UV-A BLB reactors at the same irradiance of 4 mW cm" 2 . CO 2
concentration was recorded every minute and appears to be affected by the sample stream valve
position giving two parallel trend lines.
5]5 Irradiance at the Surface of Catalyst (mW cnr 2 )
517 Figure 8. Effect of the average irradiance over the catalyst surface on STC-catalyzed PCO in
518 the LED reactor: (a) reactor effluent composition at the pseudo-steady state, PCO efficiency in
519 terms of (b) ethanol removal, (c) ethanol mineralization, and (d) PCO rate constant, and (e)
520 photonic efficiency.